Sex differences in the morbidity reported in health surveys have been noted since early in the present century. Increases in reported morbidity with age are observed, but taken for granted. The purpose of the research outlined in this proposal is to test a series of hypotheses to account for sex differences among older people, based on the proposition that these differences result, in part, from the social and psychological concomitants of gender and age identity. This research will be carried out using existing health survey data (the Health Interview Survey and the National Ambulatory Medical Care Survey) collected by the U.S. National Center for Health Statistics. Substantial social and financial costs are exacted by morbidity among women and among the elderly. Greater understanding of ways in which socially defined roles contribute to this morbidity may indicate ways of reducing some of these costs.